Geddes on Waveguides

Jump to Latest
#3,202
Michael

The waveguide you considered worse actually had the smallest mouth and aproximately the same area flare rate as the first one. Looking at the lower frequency extension, the description mentioned in the Newell & Holland book recommended by Lynn was right on.
 
#3,203

Attachments

  • beyma dome horn.jpg
    beyma dome horn.jpg
    20.7 KB · Views: 836
#3,206
Torus for full range

Please bear with me if this post is out of place here.
I've seen the good results of 4" dia. tubes at the sides of my present OB. I remember Earl made a comment awhile ago about using a Torus (imagine the shape of an inner tube) as an OB waveguide.
I wonder if using a torus for a 12" full range speaker would be a worthwhile project.
Cutting 6" dia. tubing into an 8 section makeshift torus is one possibility, another is the (temporary) use of a properly sized tire to directly attach the speaker to.
Aside from the crazy looking setup, is there any advice as to the diameter of the torus, the material used, or the cross section ?
If the results are satisfying, I'd plan on making a fiberglas Torus OB.
 
#3,207
OB is certainly off topic, but I'll comment briefly. If one wants the lowest possible diffraction from an OB, then the torus does this. However, the dogma for OB is that baffle diffraction is not an issue. I, of course, don't accept this, because there IS a lot of diffraction and I know of no situation where diffraction is a "good thing".

Now, how big should the torus be? Well simply put, the bigger the better, all the way up to an infinite baffle, which has the lowest diffraction of anything. The larger the torus the greater will be the dipole moment and the greater the LF output. It's that simple really. Not much to it.

For material, it should be dead. Foam filled fiberglas would be great. I also though that a truck tire filed with foam and then the tire cut away would be good.

Hope this helps, and I'd love to see some actual data.
 
#3,208
gedlee said:

However, the dogma for OB is that baffle diffraction is not an issue.
Click to expand...


There are good sim software to deal with this issue 😉


Anyway, I just thought about a short linearray with the cheap dome drive Beyma horn, with say 10 horns each side
Crossed to maybe 6x Beta-8a
Cost would be a problem though, so its not likely to happen:cannotbe:
Well, maybe MTM would do

What would happen if 4 of these horns were mounted in a square
 

Attachments

  • beyma dome horn.jpg
    beyma dome horn.jpg
    20.7 KB · Views: 669
#3,210
gedlee said:
OB is certainly off topic, but I'll comment briefly. If one wants the lowest possible diffraction from an OB, then the torus does this. However, the dogma for OB is that baffle diffraction is not an issue. I, of course, don't accept this, because there IS a lot of diffraction and I know of no situation where diffraction is a "good thing".

Now, how big should the torus be? Well simply put, the bigger the better, all the way up to an infinite baffle, which has the lowest diffraction of anything. The larger the torus the greater will be the dipole moment and the greater the LF output. It's that simple really. Not much to it.

For material, it should be dead. Foam filled fiberglas would be great. I also though that a truck tire filed with foam and then the tire cut away would be good.

Hope this helps, and I'd love to see some actual data.
Click to expand...


Or use a swim ring. And fill it with plaster or concrete.

http://images.google.no/images?hl=no&um=1&sa=1&q=swim+ring&aq=f&oq=
 
#3,213
I thought it would be interesting to post comparision of some simulations and the measurement.

An externally hosted image should be here but it was not working when we last tested it.


Longest curve is the measured.
Smoothest curve is simulation of waveguide in an infinit baffle.
Curve closer to the measurement dip is wave guide lip edge going straight back to an infinit baffle 1M behind.

These sims seem to indicate how the lip rolls back will effect the response.

Any views and comments are wellcome.
 
#3,214
soongsc said:
I thought it would be interesting to post comparision of some simulations and the measurement.


Longest curve is the measured.
Smoothest curve is simulation of waveguide in an infinit baffle.
Curve closer to the measurement dip is wave guide lip edge going straight back to an infinit baffle 1M behind.

These sims seem to indicate how the lip rolls back will effect the response.

Any views and comments are wellcome.
Click to expand...


Thanks for sharing!

Despite a small frequency shift that may come from slightly different dimensions in simu horn versus built horn there is remarkable tight correlation at the HF irregularity 20kHz and up.

The ripples (below 20kHz) that do not show up in infinite baffle simu are exactly what showed up as "folds in the skin" of my whale simus.
I did it with the IB of the simu set roughly at the diaphragm plain to get sort of dipole symmetry (which sadly can't mimic "real" dipole behaviour of course).

So I guess I also have to built further prototypes to measure the "real" roughness of FR - just as you did.

Michael
 
#3,215
BTW

this is my most recent attempt to bend LeCleach towards constant directivity by using the Gaussian algorithm.

DiHo_Gauss-JMLC_sf.gif



DiHo_Gauss-JMLC.gif


Resolution of simu is moderate 100 points over frequency range shown.

Comparing to the LeCleach simu shown in

http://www.diyaudio.com/forums/showthread.php?postid=1891535#post1891535

LC_IB.gif



we see considerable progress towards CD *without* compromising LeCleach' inherent sound field smootness!
🙂

If we take the yellow color as reference it is visibly more horizontal – with the result that the yellow contour now meets 90deg at roughly 2kHz - half of the 5kHz before.
On-axis and within the +/- 30deg range the overall SPL smoothness has greatly improved as well.


Very encouraging so far.

Michael
 
#3,220
gedlee said:


To be truely usefull you need to compare sims with measurements over the total polar response, not just a single axis.
Click to expand...
I quite agree, but with the limitation of geometry modeling, I doubt they will match close enough. But it's interesting to use sims to figure out what adjustments to make. Hmm, if off axis response curves can be updated in one display, that would be usefull.
 
You must log in or register to reply here.
Top Bottom